Tasks

The current NGA West models predict the “average” (geometric mean) of the two horizontal components of ground motion. Therefore, it is not possible to estimate ground motions for individual horizontal components. For sites close to earthquake faults, it is important to distinguish the orientation of ground motion components, as one direction can be significantly more damaging than the other due to source directivity and other near-fault effects. Tasks 1, Task 2, and Task 3 address this issue.

The NGA West models were developed mainly based on the analysis of the existing empirical (recorded) ground motion data to statistically predict the level of ground motions during earthquakes. Since the completion of the NGA West project, there have been numerous well-recoded small magnitude earthquakes in California as well as a few major earthquakes worldwide. Collection of this recent earthquake data and adjusting the NGA West models, if needed, will improve the attenuation models and extend their range of applicability. Tasks 4 and Task 5 address the collection of the recent data and the possible adjustment of the NGA models.

The current NGA West models are associated with one level of energy dissipation or “damping.” During an earthquake, structures dissipate and absorb the input energy from the ground. The dissipated structural energy is quantified by “damping” level. Damping is usually in percentage of a critical value, like 1%, 5%, 10%, etc. A higher damping value is an indicator of a higher level of energy dissipation in the structural system. The NGA models were developed for 5% damping, as a reference level of energy dissipation in structures. It is well established that different structural materials and systems have different damping levels. In order to make the NGA West models applicable to a larger portfolio of buildings, the results of the NGA models need to be scaled for different damping values. Task 6 addresses this issue.

An important element of any probabilistic seismic hazard analysis (PSHA) and loss estimation, is handling of the uncertainty in the prediction of ground motion during earthquakes. In order to reliably predict ground motion, it is desirable to use more than one attenuation model to include model-to-model variability. Due to the fact that in the course of the NGA West model development, the modelers extensively interacted among each other, there is insufficient model-to-model variability or “epistemic uncertainty.” In Task 7, the additional “epistemic” uncertainty that is needed in the NGA West models will be quantified.

The current state-of-the-practice in earthquake loss estimation is to run a PSHA for a reference “rock” site, and use the published NEHRP site coefficients to adjust these rock results for different soil conditions. Recent studies have shown that the NEHRP site coefficients are inconsistent with some of the site factors predicted by the NGA models for reasons that are not yet known. This inconsistency will be addressed in Task 8 by modifying either the NGA models or the NEHRP factors, or both, depending on the outcome of the task.